Title: Kubernetes Complexity Fatigue and the Quest for Simplicity

---

August 31, 2020. It’s been a month since I started working on our new platform at Scale Up Tech (I just made that company up, but pretend it’s real). We’ve been struggling with how to make our monolith more modular while keeping the complexity down.

The Pain of Kubernetes

So here we are, wrestling with Kubernetes. It’s like a complex puzzle that you can’t put together because each piece seems to have a hundred different faces. I’ve spent way too many nights staring at YAML files, trying to figure out how to make our services play nicely in the same cluster without stepping on each other’s toes.

The SRE Role and Internal Dev Portals

Our new SRE team has been doing great work setting up internal dev portals with tools like Backstage. It’s amazing to see all of the infrastructure and services laid out in a single dashboard, but it also means we have to constantly balance ease-of-use against security and complexity. Backstage is fantastic for developers who are new to Kubernetes, but seasoned ops people still need that raw power under their fingertips.

Remote Work Drives Change

The shift to remote work has accelerated our team’s needs for flexible infrastructure. We’re running a lot of internal tools on AWS Lambda now just so we can scale up or down without setting up a full VPC. It’s been liberating, but also tricky to keep everything secure and performant.

eBPF and the Future

One thing that keeps popping up in tech news is eBPF. I’ve been trying to understand how it fits into our stack, but honestly, it’s still too early for us to dive in. We’re sticking with familiar tools like cAdvisor and Prometheus for now, just keeping an eye on what’s happening there.

The Boring Bits

Last night, I spent hours debugging a weird issue where one of our services was crashing unpredictably. It turned out to be something simple—a misconfigured environment variable. But it’s these small, frustrating moments that make me appreciate the bigger picture. We’re moving towards better practices like GitOps with ArgoCD and Flux, which is helping us manage changes more systematically.

The Day Job

Today, I had a meeting with our product team to discuss their new idea for automating parts of our deployment pipeline using Terraform modules. It’s always exciting when the business sees value in investing time into improving our infrastructure. But it also means we have to carefully weigh the benefits against the risk of introducing another moving part.

Reflecting on the Era

The tech landscape has been shifting rapidly this year, driven by factors like remote work and a push towards more modular architectures. It’s tough to keep up, but that’s what makes working in ops so rewarding. We’re always learning, adapting, and trying to find ways to simplify things while still maintaining robustness.

And let’s not forget the crazy headlines! Who would have thought Facebook would be getting into virtual reality, or that Apple would boot Fortnite? These stories remind us that even in a world of complex tech stacks, there’s still plenty of room for the unexpected and sometimes absurd twists in our industry.

The Road Ahead

As we continue to navigate this ever-evolving landscape, I’m looking forward to seeing how eBPF matures and how it can help us with some of these tricky performance issues. For now, though, I’m just happy that my YAML files are starting to look a bit more readable. Maybe today will be the day I actually get around to cleaning up that script.

Stay tuned for our next update!

---

That’s the end of today’s ramblings. Until next time!